Friday, June 11, 2010

Talk to Me

I am often asked about “talking computers.” It took me a while before I understood the question was not so much about computers that talk to you (voice out) – but about computers you talk to (voice in).

The concept is pretty fabulous; just tell the computer what to do. Forget about the screen, and the keyboard you can’t see, and pay no attention to that pesky mouse. This is what we all want, isn’t it, to give a command and have it carried out?

It is voice recognition that allows us to tell our phones who to call, ask 411 to give us a number, get schedule and fare information for trains or buses. The Mac OS lets us use some voice commands, Windows has built in a speech recognition feature, and Dragon programs are perceived as pure magic. They all require serious set up, and if you have trouble reading the screen, or operating a computer, you won’t get very far.

There is one person I know – only one --who uses a computer strictly with speech recognition. His name is Ben. He is not able to see the screen and was not previously a computer user. Yes, pretty remarkable, I would agree. What he doesn’t have in visual acuity he compensates for with pure tenacity.

Ben explored the Mac, and the PC, and Dragon – only to learn that they were not solutions for Ben. He came across assistive, third-party software called Guide and with a LOT of help from his family and friends, he began using it successfully to dictate emails and get to some web pages.

While he’s been happy to join the email generation and take a peek into the World Wide Web, Guide is no nirvana. It possesses the key characteristics of assistive, third-party programs – very glitchy and very pricey.

Ben dictated an email to me yesterday, telling me that he is getting ready to replace his computer and wanted to know if there is anything new that would do a better job for him. He thought we should show Guide to Apple and perhaps we could inspire their developers, who have proven to be the best of the best, to build us a comprehensive program that is as accessible, and as simple to operate as Voice Control in the iPhone.

I told Ben everybody wants exactly what he wants, and we will get it…soon. I’m sure Apple, Microsoft, IBM and many, many others are working diligently to make our wishes come true. It’s just not quite ready…yet.

There must be someone else out there, other than Ben, who has successfully integrated accessible speech recognition into their computing.

Talk to me ---tell me what you know.
Posted by Dorrie at 6:51 PM

Labels: , , , , , , , , , ,

1 comment:

  1. UnknownJune 14, 2010 at 8:48 PM

    Hi Dorrie,

    Great blog, great topics, and great writing. It’s gratifying to hear you say you know only 1 person who uses speech in and out on a daily basis. I'm an access technology professional, and one of my roles is computer access evaluations.

    I hear a lot from people who are blind or low vision who are excited to use speech input. However, my experience is similar to yours. Speech recognition is extremely difficult and cumbersome to learn and use, especially if you are blind. I find that many people mistakenly believe speech input is "the" solution, and only after spending a lot of money and countless hours do they come to realize that it is not the solution they were hoping for. Don't get me wrong, I think Guide-Handsfree, J-Say, and similar solutions are fantastic programs, and have come a long way. They are a godsend for those who also have mobility issues where keyboarding is not an option.

    The problem, I believe, are people’s expectations of what computers can do. That is, it’s not that speech recognition is not good enough; but that computers are not human enough. Computers cannot infer or guess what we mean. They cannot know the difference between “a lite beer” and “a light bear”. It requires taking context into account, as well as tone, infection, content, etc. The way a human brain comprehends speech is seemingly impossibly complex. Yet a computer, no matter how powerful, is still a computer, bottom-up, and logical. As you know, when you give the computer a command, you have to use the right words in the right order. And learning all these commands can end up being more difficult to learn than touch-typing.

    I hope you are right that one day we can carry on a conversation with a computer just like we talk amongst are selves, but I personally think it will be a long time. And until then, learning how to touch-type is truly one of the most essential skills there are for those who are mouse-less. Just my “to sense” :)

    Cheers,
    Mark Stimson, Ph.D. (the Ph.D. is in Applied Cognitive Psychology)
    marks@accessingenuity.com

    ReplyDelete

To prevent spamming or malicious content, comments will be approved by Dorrie before they are posted.

Thanks for commenting!

Click here if you need help on submitting a comment.

Subscribe to: Post Comments (Atom)